ITB Journal of Information and Communication Technology,
Vol. 1, No. 2, November 2008. pp. 84-98. ISSN: 1978-3086
1
Experimental Study of DQPSK Modulation on SDR
Platform
1,2
Eko Marpanaji, 2Bambang Riyanto T., 2Armein Z.R. Langi, 2Adit Kurniawan,
2
Andri Mahendra, 2Thay Liung
1
Departement of Electronics Engineering, Yogyakarta State University
Karangmalang, Yogyakarta 55281, Telp. +62-274-586168, Indonesia
2
School of Electrical Engineering and Informatics, Institut Teknologi Bandung
Jl. Ganesa 10 Bandung 40132, Telp. +62-22-2501661, Fax. +62-22-2534133, Indonesia
Abstract. This Paper addresses Differential Quadriphase Shift Keying (DQPSK)
modulation implemented on SDR platform for the development of digital data
communications based on SDR. DQPSK modulation performance perceived at
Packet Error Rate (PER) is evaluated in terms of Eb/No or S/N ratio, carrier
frequency, bit rate, gain, roll-off factor of root Nyquist filter or root raised cosine
filter, and of size of payload from delivered data. Based on this results, the
smallest PER could be obtained by setting Eb/No value which is greater than 20
dB, carrier frequency of at least 0,3 MHz, optimum bit rate of 200 kbps,
optimum range payload size of 2000 up to 4000 bytes, and roll-off factor of
Nyquist or root-raised cosine filter of 0.4.
Keywords: Differential quadriphase shift keying, packet error rate, root raised cosine,
software-defined radio, usrp.
1
Introduction
This paper describes the implementation of Differential Quadriphase Shift
Keying (DQPSK) modulation scheme for communicating of digital data on
Software-Defined Radio (SDR) platform. This performance evaluation of
DQPSK modulation scheme is also as a part of the current research in building
SDR platform which can be reconfigured, including answering the research
question of “How to design SDR platform using commodity components.”
Chen (2007) investigated a performance of Quadriphase Shift Keying (QPSK)
modulation scheme to send a video digital data. All algorithms are modeled and
simulated in Matlab, and the future works of that research is integrated those
model into GNU Radio SDR platform using USRP[1]. Thus, the investigation
of QPSK modulation scheme performance and its variants, including DQPSK,
on SDR platform using Universal Software Radio Peripheral (USRP) is still
needed.
Received ________, Revised _________, Accepted for publication __________
( Makalah diterima redaksi tanggal _______, revisi diterima tanggal ________, diterbitkan tanggal _______ )
2
Author’s name
In this research, we investigated the performance of DQPSK modulation
scheme implemented on SDR platform using USRP from GNU Radio. DQPSK
modulation scheme is chosen in this research because the QPSK and DQPSK
modulation is also widely used in recent wireless and cellular communication
systems such as Satellite, CDMA, and cable modem.
The rest of this paper is organized as follows. Section 2 discusses about SDR
architecture along with SDR’s characteristics. Section 3 presents the DQPSK
modulation scheme. Section 4 presents system configuration used in this
research. Section 5 presents DQPSK implementation and test result. Finally,
Section 6 is conclusion.
2
Software-Defined Radio (SDR)
Software-defined radio (SDR), sometimes shortened to software radio (SR),
was introduced for the first time in 1991 by Joseph Mitola[2]. The word of SDR
was used to show a radio class that could be re-configured or reprogrammed[3], thus resulted a kind application of wireless communication
with mode and frequency band determined by software function. Ideally, SDR
offers flexibility, re-configurability, scalability and as multi mode as possible.
SDR architecture is developed based on conventional radio functions. The
difference is all functions of signal processing on conventional radio are carried
out fully by hardware while the functions of signal processing on SDR are
carried out as much as possible by software. The major key in building SDR is
the placement of ADC and DAC components as a divider between analog and
digital domain, thus the signal processing can be carried out using software.
Wideband
Antenna
Wideband
Analog IF
Signal
RF Front End
(Up/Down
Converter, RF
Amplifier
Wideband
ADC/DAC
ADC
DAC
Processor
and
Memory
Data
Out
Data
In
Figure 1Realistic SDR Architecture
The ideal SDR architecture will place ADC/DAC as close as possible with the
antenna, and will be followed by processor to carry out digital signal
processing. The radio functions are carried out by the software which is
operated by the current processor. Therefore, if there is standard change of
Title of Paper (11 pt, Century Gothic, max. 50 character)
3
communication, thus its change can be anticipated by upgrading the
software[2][3]. However, the technology limitation and the expensive
ADC/DAC wideband urge to change a little of SDR architecture in placing
ADC/DAC. The more realistic SDR architecture places ADC/DAC wideband
after Down Converter/Up Converter as shown in Fig. 1. With this architecture,
the conversion from analog to digital or reverse is carried out on Intermediate
Frequency (IF) signal which possesses lower frequency than RF signal. Today,
that type of architecture are being developed widely and researched for the
implementation. The SDR architecture for both transmitter and receiver can be
presented with block diagram as shown in Fig. 2.
Receiver (Rx)
Software
Wideband
Antenna
Channelization
and Digital Down
Converter
Wideband RF
Front End
(RF Amp. and Up/
Down Conv.)
Analog Domain
IF Signal
Processing
(Gain, AGC)
Demodulation
ADC
and
DAC
Processor
and
Memory
Hardware
Baseband
Signal
Processing
Digital
Data
(bitstream)
Digital Domain
IF Digital
Signal
Processing
Modulation
Baseband
Signal
Processing
Software
Transmitter (Tx)
Figure 2SDR Architecture for Transmitter and Receiver
The SDR platform performs transmitting and receiving functions. The
transmitter (Tx) will perform some process such as base band signal processing,
modulation, digital IF signal processing, and sending RF signal to the air. The
receiver (Rx) will perform some process such as RF signal processing,
channelization, digital IF signal processing, demodulation, and base band signal
processing. As shown in Fig. 2, the computation process in the receiver will be
more complex than the transmitter.
ADC and DAC component will determine a point in which radio functions can
be performed through a software. This point is usually called digital access
point. The realistic SDR architecture will use software to process the digital IF
signal. The major challenge in choosing ADC’s location is the problem of
limited sampling frequency and economics factor. If the sampling frequency of
Nyquist is hard to be done because of its digitalization IF signal is heavy, then
the alternative is to apply under sampling toward IF signal, therefore it reduces
4
Author’s name
the computation complexity from processor. The formula to determine under
sampling frequency is as shown below:
2 fc + B
2 fc − B
≤ fs ≤
;m∈ N
m+ 1
m
(1)
Where f C is center frequency, B is bandwidth, f s is sampling frequency
chosen until f s > 2 B [4]. For example, the SDR system uses 20 MHz IF signal
with 5 MHz bandwidth can be sampled with frequency of 22,5 MHz for m=1, or
17,5 MHz for m=2, or 11,66 MHz for m=3.
The next attention is on computation complexity which will determine the type
and size of processor needed. The processor will generate the software of digital
signal processing both for the receiver and transmitter. Besides to observe the
functions generated by the software in determining computation complexity, the
scalability factor for next development anticipation is also observed. All those
steps will determine the processor’s size usually stated with Million Operations
per Second (MOPS) or Million Instructions per Second (MIPS).
The biggest computation requirement is on the receiver side especially on filter
part in channelization, which range from 100 to 200 operation/sample[3][5][6].
By using the sampling speed of 22.5 MHz will result 22.5 MSPS, thus it needs
2,250 until 4,500 MIPS. Other computation requirements are modulation and
demodulation process. As a description, modulation and demodulation FM
require more or less 100 to 150 MIPS[7]. The total of these computation
requirements will determine the type and the number of processor used in SDR
platform. It means, for bigger computation requirement and unable to be carried
out by a processor, thus the SDR architecture can use parallel computation
architecture by involving several processors or implementing a part of signal
processing function in form of FPGA.
In the next section, the SDR implementation for digital data communication
using Differential Quadriphase Shift Keying (DQPSK) modulation scheme
along with the test result to observe its PER will be explained.
3
Differential Quadriphase Shift Keying (DQPSK) Modulation
DQPSK modulation is a variant of π 4 -QPSK modulation scheme, which
differential encoding is added to the bits encoding prior to the modulation.
Differential means that the information is not carried by the absolute state; it is
carried by the transition between states. The advantage of using differential
encoding is free from phase ambiguity if the constellation is rotated by some
Title of Paper (11 pt, Century Gothic, max. 50 character)
5
effect in the communications channel which the signal passes through. This
problem can be overcome by using the data to change rather than set the phase.
The QPSK modulation scheme using differential encoding is usually called as
π 4 -DQPSK or DQPSK.
The equation differential encoder to produce symbols is:
s k = s k − 1 ⊕ bk
(2)
where s k is current output symbol, s k − 1 is previous output symbol, and bk is
current bit input. The ⊕ sign indicate modulo-2 operation. So s k only changes
state (from binary '0' to binary '1' or from binary '1' to binary '0') if bk is a
binary '1'. Otherwise it remains in its previous state. At the receiver, the
received signal is demodulated to yield s k = ± 1 and then differential decoder
reverses the encoding procedure (decoding process) and produces:
b′ = s k ⊕ s k − 1
(3)
since binary subtraction is the same as binary addition. Therefore, bk′ = 1 if s k
and s k − 1 differ and bk′ = 0 if they are the same. Hence, if both s k and s k − 1 are
inverted, bk will still be decoded correctly as bk′ . Thus, the 180° phase
ambiguity does not matter.
The DQPSK modulation format uses two QPSK constellations offset by 45
degrees (π/4 radians) as depicted in Fig. 3. Transitions must occur from one
constellation to the other, or each DQPSK symbol will reside in one of eight
points in the constellation diagram. This guarantees that there is always a
change in phase at each symbol, making clock recovery easier. The data is
encoded in the magnitude and direction of the phase shift, not in the absolute
position on the constellation.
6
Author’s name
Q
01
01
11
00
45o
11
00
I
10
10
Figure 3DPSK constellation diagram with identical Gray coding rotated by 45o
Table 1 summarizes a possible set of relationships between the phase transitions
in the DQPSK modulation scheme and the incoming Gray code dibits[8].
Table 1Correspondence between input dibit and phase change for DQPSK
modulation
Gray-Encoded Input
Dibit
Phase change,
∆ θ (radians)
00
π 4
01
3π 4
11
− 3π 4
10
−π 4
DQPSK used root Nyquist filter or root raised cosine filter as a pulse-shaping
filter. This filter is used to reduces a number of spurious signals or to control
the shape of digital data. The impulse response of root Nyquist filter is
determined by equation [3]:
g (t ) =
1
π t
1
 4α t 

1 − 

 Tb 
2

1 1
sin  2π (1 − α )  +
Tb  π

4 α / Tb
 4α t 

1 − 

 Tb 
2

1
cos  2π (1 + α ) 
Tb 

(4)
Title of Paper (11 pt, Century Gothic, max. 50 character)
7
where α is roll-off factor of root raised cosine filter. This parameter controls
the shape and bandwidth of the incoming signal.
N y q u ist p u ls e s h a p in g -tim e a x is
1
0 .8
0 .6
a m p litu d e
a lp h a = 0
0 .4
a lp h a = 0 .5
a lp h a = 1
0 .2
0
-0 .2
-0 .4
-3 T b
-2 T b
- Tb
0
tim e
Tb
2Tb
3Tb
Figure 4Impulse response of a Nyquist filter in time domain with several
chosen roll-off factors.
Fig. 4 shows the impulse response of a root Nyquist filter in time domain with
several chosen of roll-off factor (alpha) values. One features of the Nyquist
filter is that we always obtain impulse response is zero at nTb (n is integer: 1,
2, 3, ...) in the time domain. Therefore, when we set the synchronization point at
nTb , one symbol never interferes with other symbols at this point.
The advantages of DQPSK are:
−
the phase transitions from one symbol to the next are restricted to ± π 4
and ± 3π 4 so it will reduce the symbol ambiguity
−
using root raised cosine filtering, it has better spectral efficiency than
Gaussian Minimum Shift Keying (GMSK), the other common cellular
modulation type.
8
Author’s name
3.1
DQPSK Modulator
Fig. 5 shows the principle of a DQPSK modulator. The process occurred on
modulator starts by changing bit stream with level 0 and 1 into bipolar stream
bit with level -1 and 1. Next, bipolar stream bit is processed by mapping circuit
to produce dibits for I-channel (Ich) and Q-channel (Qch), and then each dibit
data will be processed by differential encoder using the rules of Table 1
separately. The output of differential encoder will be passed through a pulseshaping filter to reduce its spurious signals. Finally, the output of pulse-shaping
filter is converted to analog signal and modulated using RF signal cos(2π f C t )
for Ich signal and sin(2π f C t ) for Qch signal, where f C is carrier frequency.
Both Ich and Qch signal are combine together to produce DQPSK signal.
Diff.
Encoder
Ich Data
PulseShaping
Filter
D/A
Conv
cos(2π f C t )
Bit stream
[0,1] to
[-1,1]
Mapping
Circuit
DQPSK
signal
sin( 2π f C t )
Qch Data
Diff.
Encoder
PulseShaping
Filter
D/A
Conv
Figure 5DQPSK Modulator
3.2
DQPSK Demodulator
At receiver or demodulator of DQPSK, process that occurred is reverse at
modulator of DQPSK. Fig 6 showing block diagram of DQPSK demodulator.
Title of Paper (11 pt, Century Gothic, max. 50 character)
A/D
Conv
PulseShaping
Filter
9
Diff.
Decoder
Ich Data
DQPSK
signal
cos(2π f C t + θ 1 (t ))
Demapping
Circuit
[-1,1]
to
[0,1]
Bit
stream
sin(2π f C t + θ 1 (t ))
Qch Data
A/D
Conv
PulseShaping
Filter
Diff.
Decoder
Figure 6DQPSK Demodulator
Referred to the figure, originally demodulator of DQPSK will eliminate carrier
frequency so that recover Ich and Qch signal. Both signal then converted in the
form of digital signal through A/D Converter. Digital signal which yielded then
passes through a pulse-shaping filter, and by using differential decoder dibit
information transmitted by modulator will be recovered. Dibit information will
be dissociated and become consecutive bit through demapping circuit, and bit
stream with level - 1 and 1 will be returned to bit stream with level 0 and 1.
Thus, consecutive information bit data transmitted by modulator of DQPSK can
be found again at DQPSK demodulator.
4
System Configuration
System configuration which used in this research is shown at Fig 7. SDR
platform in this research use USRP GNU Radio peripheral that function is
up/down converter as a front end and personal computer (PC) as the processor.
System configuration of SDR using PC as the processor was often referred as
Software Radio or shortened by SWR. Pursuant to configuration, hence
modulator and demodulator of DQPSK can be implemented in software, while
A /D and D/A converter remain in hardware form which located in USRP.
USRP and PC is connected using port USB version 2.0.
10
Author’s name
Tx
Rx
DQPSK
Modulator
digital data
(bit stream)
Up
Converter
Down
Converter
Hardware
Software
DQPSK
Demodulator
Hardware
digital data
(bit stream)
Software
Transmission
Line/
AWGN channel
USRP
Computer
USRP
OS : Linux
Python Programming
Computer
OS : Linux
Python Programming
Figure 7System configuration of SDR platform
Mainboard specification of USRP which used in this research is: (a) USB 2.0
port for connection to computer; (b) 12-bit ADC with speed of sampling 64
MSPS so that with principle of aliasing can conduct digitalization process with
range frequency of aliasing - 32 MHz until 32 MHz; (c) 14-bit DAC with clock
frequency 128 MSPS so have Nyquist frequency equal to 64 MHz; and (d)
analog signal which yielded limited to 10 mWatt. While daughterboard using
Basic Tx and Basic Rx so that there is no process of up/down converter and
transmitter frequency limited 50 MHz maximum.
Specification of PC which used in this research is summarized at Table 2.
Computer that used must have USB version 2.0 port to support connection with
USRP board. Programming language used Phyton and a few block functions
written in C++.
Table 2Computer spesifications that used in this research
No
Component
1.
Processor
2.
RAM
3.
Operating
System
1st Computer
AMD Athlon
XP1800+, 1.53
GHz, fsb 533
MHz
DDR 333 MHz,
256 Mbyte
Linux Fedora
Core-4, 2.6.111.i369
2nd Computer
Intel Pentium
4, 2.93 Ghz,
fsb 533 MHz
DDR
400
MHz, 2 x 256
MByte
Linux Fedora
Core-4,
2.6.11-1.i369
Title of Paper (11 pt, Century Gothic, max. 50 character)
11
System performance can be measured using one of the parameter which was
often used in QoS, that is:
BER < 10-3, PLR or PER < 10-2, spread delay < 100 ms, and Gos > 95% [4].
In this research performance modulation scheme of DQPSK is measured in
Packet Error Rate (PER) because data transmitted in the form of packet.
Transmitting data as a packet will facilitate of SDR software integration with
TCP/IP protocol for next development.
5
Implementation and Test Result
DQPSK modulator and demodulator described previously are implemented in
SDR platform using PC and USRP board that commonly used in SDR
experiment from GNURadio. The Board is a front side of SDR system and has
functions to execute digitalization function (ADC-DAC) and filter
channelization process. The digital signal processing used personal computer
with USB port as a link between USRP and PC. Operating system used Linux
with python programming language to develop the application. All software
used for digital signal processing was open source from GNU Radio.
Figure 8Spectrum DQPSK signal
This experiment would observe the performance of digital communication
based on SDR system using USRP board with DQPSK modulation scheme. The
transmission channel of coaxial cable RG58 was used to connect transmitter and
12
Author’s name
receiver because of limitation of Basic TX output power. The experiment was
carried out to observe Packet Error Rate (PER) stated by the percentage of
exactly accepted packet toward variation of Eb/No, bit rate, carrier frequency,
roll-off factor, gain, and also payload size.
Fig. 8 shows the spectrum of DQPSK signal which has 15 MHz carrier
frequency, bit rate 256 kbps, and roll-off factor 0.35. The result of observation
is shown in Fig. 9 up to Fig. 14.
Figure 9The effect of Eb/No on PER
Fig. 9 shows the graph between variation of Eb/No concerning the PER value
with 128 kbps, 256 kbps, and 512 kbps bitrate. Referred to the graph perceived,
greater value of Eb/No hence smaller PER will be obtained. PER value will
remain low for the value of Eb/No higher than 20 dB.
Variation of frequency carrier modulations of DQPSK will affect to PER value
as depicted in Fig 10. The Graph obtained by arranged bit-rate transmitted data
equal to 128 kbps, 256 kbps, and 500 kbps. In this experiment maximum
frequency that tried is 44 MHz, and beyond that frequency cannot be handled
by USRP using existing Basic TX. Referred to the graph, lowest PER value can
be obtained if f C > 0.06 MHz for 128 kbps, f C > 0.2 MHz for 256 kbps, and
f C > 0.3 MHz for 500 kbps. In another word, the minimum frequency which
support for transmitting data with bit-rate greater than 256 kbps is 0.3 MHz.
Title of Paper (11 pt, Century Gothic, max. 50 character)
13
Figure 10The effect of carrier frequency on PER
Figure 11The effect of bit rate on PER
Fig. 11 shows a graph of bit-rate variation to PER value. Referred to the graph,
the optimum bit-rate for the modulation of this DQPSK is 200 kbps. Bit rate
14
Author’s name
greater than 200 kbps will increase the PER value and bit rate from 250 up to
600 kbps produced higher PER value.
Fig. 12 shows a graph of roll-off factor variation value at Nyquist filter to PER
value. Referred to the graph, in general can be expressed that bigger roll-off
factor value will result smaller PER, and optimum value for roll-off factor to get
PER is 0.4. Roll-off factor value for the Nyquist of filter is 0 ≤ α ≤ 1, which
value of roll-off equal to 1 referred as roll-off full-cosine which is very useful
in signal timing extraction for the process of synchronization. Ever greater
value of roll-off factor hence smaller influence of ISI.
Figure 12The effect of roll-off factor on PER
Fig. 13 shows a graph between the variation of gain value and resulted PER
value. Referred to the graph, the greater gain will result the smaller PER value.
However, for the gain value bigger than 500 PER value do not change the PER
value. Hence, the optimum value for gain is 500.
Title of Paper (11 pt, Century Gothic, max. 50 character)
15
Figure 13The effect of gain PER
Fig 14 shows a graph concerning with variation of payload value to PER value.
The graph indicated that the bigger transmitted payload value will produce the
smaller PER. The maximum data size for this system is 4092 bytes, hence the
optimum value for the measure of payload is 4092 bytes.
Figure 14The effect of payload size on PER
16
Author’s name
Referred to all the graphs of this research result, PER value which obtained with
variation of Eb/No, carrier frequency, bit rate, roll-off factor, gain, and payload
is very low that is closing 10-2. In spite of comparison with criteria measurement
performance where PER < 10-2 not yet fulfilled. Hence, modulation scheme
implementation of DQPSK at platform of SDR use USRP of GNU Radio still
require to be completed either software and hardware so that we can get better
performance of DQPSK modulation scheme on SDR platform using USRP
GNU Radio.
6
Conclusion
Referred to this research result we can conclude that: In general can be said that
modulation scheme of DQPSK which implemented at SDR platform using
USRP from GNU Radio have better performance which is seen from PER
value. Lowest PER Value can be obtained at Eb/No value > 20 dB, optimum
bit-rate 200 kbps, carrier frequency bigger than 0,3 MHz, roll-off factor (alpha)
is 0.4, optimum gain is 500, and optimum payload value is 4092 byte.
Continuation of this research is software development of SDR to improve the
performance and also create application that adept to integration with
communications protocol for example TCP/IP.
7
Acknowledgment
We would like to thank to the Dean of STEI ITB and the chief of PPTIK STEI
ITB who have given opportunity also donation in order to accomplish this
research. This paper is a part of the current research which has been conducting
about SDR platform development, and it is using ITB Research fund year 2007
No : 0047/K01.03.2/PL2.1.5/I/2007. We also thank to all people who contribute
to this research.
8
[1]
[2]
[3]
[4]
References
Z. Chen., “Performance analysis of channel estimation and adaptive
equalization in slow fading channel,” 2007, [On-line], Available at
http://www.wu.ece.ufl.edu/projects/channelEstima-tion/full_files.zip.
J. H. Reed, Software Radio: A Modern Approach to Radio Engineering.
New Jersey: Prentice Hall, 2002, pp. 2-3.
J. Mitola III, Software Radio Architecture. Object-Oriented Approaches
to Wireless Systems Engineering. Canada: John Eiley & Sons, Inc, 2000,
pp. 1-34.
M. E. Angoletta. (2003, Feb. 20). From Analog to Digital Domain.
[Online]. Available: http://humanresources.web.cern.ch/ humanre-
Title of Paper (11 pt, Century Gothic, max. 50 character)
17
sources/external/training/special/
DISP2003/DISP-2003_L01A_20Feb
03.pdf (20 Feb 2003).
[5] A. K. Salkintzis, H. Nie, P. T. Mathiooulos. (1999, August). ADC and
DSP Challenges in the Development of Software Radio Base Stations.
IEEE Personal Communications.
[6] C. H. Dick. (2001). Design and Implementation of High-Performance
FPGA Signal Processing Datapath for Software Define Radios. [Online].
Available:
http://
www.eetasia.com/ARTICLES/2001AUG/
2001AUG09_ ICD_AMD_RFD_ TAC01 .PDF.
[7] J. Huie, P. D’Antonio, R. Pelt. (2005). Synthesizing FPGA Cores for
Software-Defined Radio. [Online]. Available: http://www.altera.com/
cp/fpga-cores-for-sdr.pdf.
[8] S. Haykin, “Communication System,” 4th ed. New York: John Wiley &
Sons, Inc, 2004, pp. 362-365.
[9] H. Harada, R. Prasad, Simulation and Software Radio for Mobile
Communications. Boston: Artech House, 2002, pp. 75-76.
[10] J. Mitola III, Software Radio Architecture. Object-Oriented Approaches
to Wireless Systems Engineering. Canada: John Eiley & Sons, Inc, 2000,
pp. 165-166.